INDEX

Explanations

words related to forceful and explosive actions

oai_token-act-pair · gpt-3.5-turbo

references to explosive events or phenomena

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PN

-0.75

phis

-0.74

tymology

-0.74

ccording

-0.72

abet

-0.71

 compr

-0.70

sit

-0.70

Adapt

-0.69

guyen

-0.68

perate

-0.68

POSITIVE LOGITS

 furnace

1.01

 blast

0.96

blast

0.96

 blasts

0.88

asted

0.86

waves

0.85

 Heard

0.84

ographed

0.83

ocalypse

0.83

asting

0.81

Activations Density 0.015%

words related to forceful and explosive actions

references to explosive events or phenomena

No Comments

No Known Activations